Zambuto/StarMaster quality claims
Ron Wodaski
somewhere, will manage to misunderstand you.
<sigh>
Ron
"Chris Adamson" <adamso...@home.com> wrote in message
news:Bn4q4.22774$G91.2...@news1.rdc1.az.home.com...
> I disagree. The claims are based on the specific method mentioned in the
> paperwork which is verifiable using the same method. At that same website
> (http://www.icstars.com/starmaster/ZambutoCert.html), Mr. Zambuto writes
...
>
> ------------------------------------------------------------------------
> "Information Regarding Test Method Accuracy
>
>
> Test accuracy
> The zonal Foucault test is generally considered to be accurate to 1/10
> wavefront. As applied to the amateur astronomy community in this context
> that is a fair assumption, because there are many factors involved in the
> test. Some of those factors include test setup and design of test
> parameters, skill of the operator, environmental conditions, data
reduction
> software used, production procedure and size of optic tested. Also, much
> about the procedure and execution is subjective to the operator. For those
> reasons it is appropriate to consider final numbers beyond 1/10 wavefront
as
> approximations.
>
> Comparing methods and results
> The data reduction provided on the certification should be compared with
> same test method only, that being zonal Foucault. Direct comparisons
between
> differing methods of test cannot necessarily be made. The wavefront rating
> of a zonal measurement of sample points along the radius will not be the
> same as the wavefront rating of a system which uses auxiliary optics to
> produce a reference wavefront, or a system which utilizes auxiliary optics
> to produce a null. The reasons are many, and I wonít go into them here, as
> that is not the purpose of this information. What I will say is the method
> used is an established standard in its own right, (it does not need to be
> compared with other methods) and is repeatable and verifiable using the
> information given on the document. The particular method was first
published
> in 1919. The stated criteria by Jean Texereau is as follows, quoting from
> How to Make a Telescope, Second Edition, page 105: "At the SociÈtÈ
> Astronomique, we release a mirror only when the final values of lambda*f/p
> are less than 1 and the wavefront error is under 1/10 wavelength". In a
> footnote to that statement it is further indicated, "Average wavefront
error
> on the first 18 mirrors was lambda/20.4". That is in fact the criteria
used
> by Zambuto Optical Company for 8" and 10" mirrors, at 1/20 wavefront
> (combined with an RTA of less than 5). Larger mirrors are allowed more
room
> in the diffraction envelope, but not to exceed the diffraction limit (RTA
> 1.0) combined with a wavefront error of less than 1/10, which is
Texereau's
> criteria.
>
> Repeatability
> There have only been a few customers in the two years we have been in
> commercial business who have bench tested the mirrors. In those cases, the
> results have been remarkably similar, with the total wavefront error for
the
> average of measurements taken not differing more than 1/50 wavefront (.02
> wave) from the documentation. I do not expect this to always be the case,
> especially because these were smaller mirrors tested, such as 10" and
> smaller. What I do expect however, is the deviation should not exceed more
> than about .05 or 1/20 wavefront over the entire optic. That kind of error
> between test situations would justify the estimate of 1/10 wave accuracy.
I
> encourage anyone who wants to bench test an optic to do so, as I always
> appreciate feedback regarding the mirror, the method used and etc. The
> information provided on the certificate details the zones used and all
other
> pertinent information. Furthermore if you are wanting to precisely
duplicate
> the test as I have performed it, I would be happy to send you a template
of
> the mask used, some brief information regarding orientation of the mirror,
> and raw data sets taken.
>
> The guarantee
> I often observe people stating final values of their mirrors in public
> forums. First, It should always be stated along with the test method, that
> being zonal Foucault, if it is going to be stated at all. I personally
would
> rather it were not stated, as one, it causes confusion with other methods
to
> the uninformed, and two, the numbers are not included in my actual
> guarantee. What I guarantee is found at the bottom of the document for
each
> mirror, accompanied by my signature. I guarantee performance at the
> eyepiece, (where it really matters) to beyond 50 power per inch with no
> breakdown in the in-focus image under appropriate conditions. Although I
> have observed much debate over numbers on paper over the years, I have yet
> to observe argument at the eyepiece when an optic is a superb performer.
So
> I would rather it were stated that a customer owns a "50x per inch mirror"
> than an "x-wavefront" mirror. Besides, once the RTA is less than 1.0
> accompanied by a smooth surface and a good edge, the difference at the
> eyepiece in focus is negligible.
>
> Why documentation at all?
>
> The purpose is fourfold: One, it demonstrates the method of test used,
> including all necessary information to repeat the test by anyone with a
> working knowledge of Foucault and software that is readily available. Two,
> it
> demonstrates to what criteria I work a mirror using the method described
> before I stop, or more important, to what point I work an optic to be able
> to make the 50x per inch guarantee.Three, it identifies and authenticates
> each mirror, as the accompanying document has recorded on it the mirrorís
> measurements, serial number, and my signature. And four, and perhaps most
> interesting is the document for your mirror is the actual specific source
I
> use to perform the final figuring. In active spreadsheet mode, by
> manipulating the focus point mathematically I observe the positions of
zones
> in the envelope, and use the wavefront graph itself to do the
fine-touching.
> The hills and valleys on that graph are real (as far as the test accuracy
> goes), and I actually orient lap sizes and positions with respect to that
> surface until it meets specification. So the document you have is the
final
> result of the work done in the shop on your mirror. It is that functional.
>
> Final word
> The points I want to be understood regarding documentation are three:
>
> 1. It should be understood that the numbers on the documentation represent
> the results of the zonal Foucault test as performed by me, in accordance
> with
> the test description and parameters outlined in the explanation page
> accompanying each document. That is all it represents, and nothing more.
>
> 2. Because of the nature of the test itself and the previously mentioned
> factors, values beyond 1/10 wavefront should not be taken as ìabsolute
> reality", but are better understood as an approximation.
>
> 3. The results of this test method should not be compared or confused with
> the results of any other method, and furthermore any time the results are
> stated, the method used should be clarified in the same statement, and
that
> is zonal Foucault."
>
>
> Please read the whole page at the link to get this in context. Like many
> others who have responded, happiness is in the eye of the beholder and we
> have yet to hear of one Zambuto optic owner who is unhappy.
> Regards,
> Chris
>
>
>
>
> Blandp1 <bla...@aol.com> wrote in message
> news:20000214192316...@nso-fw.aol.com...
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Just to set the record straight on what is being claimed by and
> > for Zambuto mirrors. Taken from the StarMaster web site
> > (http://www.icstars.com/starmaster/planetary.htm):
> >
> > >2.Peak to Valley wavefront measurements will be between
> > >1/15 and 1/20 wavefront through 16", and in some cases
> > >slightly better (typically in the smaller sizes such as 12.5").
> > >18 F/4.3 mirrors will be at least 1/12 P.V. at the wavefront.
> >
> > Additionally an image of a Zambuto produced test report of a
> > 14.5" mirror is shown which claims a Strehl ratio of 98.3%, with
> > P-V wavefront of 1/18.8 and RMS of 1/48. Carl Zambuto at the
> > bottom of the document guarantees 50X per inch.
> >
> > As can be seen Zambuto *is* making the claim of 1/18.8 P-V on at
> > least this mirror, and certainly StarMaster is making the claim
> > of at least 1/15 P-V on all (< 16") mirrors.
> >
> > I have no way of verifying the wavefront claims of Zambuto short
> > of buying a mirror and paying for such testing to be done. I
> > have no doubt that Zambuto makes a fine mirror, and if I where
> > to get a dob I would probably buy one with a Zambuto mirror.
> > But to say that Zambuto and/or his representatives don't make
> > P-V/Strehl claims is demonstrably false.
> >
> > Philip J. Blanda III
> >
> > -----BEGIN PGP SIGNATURE-----
> > Version: PGPfreeware 6.5.1 for non-commercial use <http://www.pgp.com>
> >
> > iQA/AwUBOKicU8qM3UPhf8W7EQL++wCeIygRC1v8TaRZc0a/4vMTd02E7rMAn1Jk
> > 3p9zeseWk2EdsGXX/q/1JxF/
> > =er7w
> > -----END PGP SIGNATURE-----
> >
> >
> >
> >
> >
> > PGP Public key available
>
>
Ron Wodaski
discourages folks from reading too much into them. If they work for him (and
apparently they do), great.
Folks can discuss the fine points all they want; discriminating buyers will
figure it out.
Ron
"Chris Adamson" <adamso...@home.com> wrote in message
news:3ppq4.23956$G91.2...@news1.rdc1.az.home.com...
> Proving only that, no matter how clearly something is stated by a person
who
> actually produces excellent optics for a living, someone, somewhere will
> manage to 1) misunderstand what is written or 2) not bother to read or 3)
> mistake their opinion for a natural law. :)
>
> In addition, I absolutely disagree with those who seem to be telescope
Ralph
> Nadar's worried about people buying high end scopes basing their decision
on
> some 1/18 claim versus someone else's 1/8, 1/4 or whatever. There are not
> that many people that will drop $4k - 8k on a scope without doing some
> degree of research beyond some marketing literature. Those who do not
> deserve what they get. StarMaster and Zambuto get sales because they have
> become well known and well respected for the quality of their product. It
> is that simple. I have yet to hear of a single individual, who actually
> purchased Zambuto optics, question the actual quality or quality claims
> (perhaps because they are always told the methods used to any level of
> detail they desire) and certainly because they are very happy with what
they
> get. That is what matters after all.
> Regards,
> Chris
>
> Ron Wodaski <rwod...@hotmail.com> wrote in message
> news:38a8...@news.nwlink.com...
PhotoKing
manufacturer about the quality of their optics. After he told me how
wonderful his optics were, I asked the manufacturer how his mirrors compared
to Zambutto mirrors. He proceeded to tell me a bunch of nonsense that he
didn't know what kind of technical expertise Zambutto has and went a long
way to impliedly denigrate his mirrors. What it demonstrated to me is the
amount of professional jealousy going on out there. Carl Zambutto has
developed a reputation that places his mirrors in the top tier for amateur
astronomers. I suggest that competitors would be better off emulating his
high quality work rather than trying to tear down his hard-earned reputation
for producing superb mirrors.
Chris Adamson <adamso...@home.com> wrote in message
Ratboy99
>discourages folks from reading too much into them.
Excellent point. I think the numbers give Carl a consistent benchmark to use
for figuring the mirrors. They are more for his use than for ours.
rat
~( );>
bratis...@my-deja.com
ratb...@aol.comet (Ratboy99) wrote:
> >I get the sense that Zambuto uses his test to perfect his mirrors,
and
> >discourages folks from reading too much into them.
>
to use
> for figuring the mirrors. They are more for his use than for ours.
So why doesn't he keep 'his' numbers for himself instead giving them to
every Joe to brag about his "1/29.3" wave mirror ? Roland Christen DOES
have all the numbers, and real ones too (off the interferometer) for his
APOs and he does indeed keep them for himself only.
I really think Zambuto and John Hall are shooting themselves in the foot
by supplying those (largely imaginary) numbers with mirrors they make.
Bratislav
Sent via Deja.com http://www.deja.com/
Before you buy.
valery_...@my-deja.com
> >discourages folks from reading too much into them.
>
> for figuring the mirrors. They are more for his use than for ours.
>
> ~( );>
Fine!
Let him keep this ONLY for himself, not for us. Why he let us
know the info which is only for his private use?
Valery.
Mike Barrs
news:88fv38$e56$1...@nnrp1.deja.com...
> and
> > >discourages folks from reading too much into them.
> >
> to use
> > for figuring the mirrors. They are more for his use than for ours.
>
> every Joe to brag about his "1/29.3" wave mirror ? Roland Christen DOES
> have all the numbers, and real ones too (off the interferometer) for his
> APOs and he does indeed keep them for himself only.
> I really think Zambuto and John Hall are shooting themselves in the foot
> by supplying those (largely imaginary) numbers with mirrors they make.
>
> Bratislav
John Hall supplies two sets of numbers for Strehl ratio, P-V wavefront, and
RMS wavefront. One set is called "indicated error, uncorrected for the
effect of random variables", and the other set is called "corrected to
simulate potential effect of random variables". That's an interesting
marketing approach. It gives you a choice of being realistic when talking
about your optics, or else using inflated numbers for bragging rights. :-)
One recent large Pegasus mirror I saw had 1/8 wavefront P-V as the corrected
number, and 1/13 wavefront P-V as the uncorrected number.
During the past two weeks, I've been getting some quotes from a few
professional telescope optics suppliers, for a 22"-24" Newtonian mirror.
I've been trying to see what it would cost to get better quality than the
typical large amateur mirror from Pegasus, Nova, or Galaxy. It's a real
eye-opener. Quotes for figuring 1/8 - 1/10 wavefront are running from $6,000
to $13,000, and that's not including the cost of the blank! It does make one
just a wee bit suspicious about the wavefront claims for amateur mirrors
costing 1/3 to 1/4 as much.
Why is it so expensive to get a mirror from these pro optical shops? Well,
for one thing, their clients are mostly government agencies and
universities. These outfits either have their own testing instruments to
confirm quality, or else they subcontract for independent testing of the
delivered product. That keeps the wavefront numbers honest. Independent
testing and confirmation is the missing ingredient in the amateur mirror
market.
Amateur mirror suppliers deliver a very good product for the money, most of
the time. But this test report numbers game is confusing, and in some cases
it borders on being dishonest (I know of two recent mirrors from a highly
reputable amateur supplier that failed a star test, by a wide margin, and
both mirrors had excellent test report numbers).
An outfit like Zambuto, Pegasus, Galaxy, or Nova should simply give a
money-back satisfaction guarantee that a new mirror will pass a star test,
period. Test report numbers are worthless, unless we have the ability to
duplicate and confirm the numbers. The star test is the amateur's equivalent
of a Zygo interferometer. It's the only realistic certification procedure we
have.
Mike Barrs
Andy Wallace
on each mirror he delivers. When I gave those numbers on my 18" mirror to
Thomas Back of this group, Thomas's comment was "Yes, these are very good
numbers; 1/6.45 P-V. And what is a realistic wavefront for an excellent
large mirror." The interferometer specs are: "over 2892 pts, peak of 0.060,
valley of -0.095, P-V of 0.155, RMS of 0.024 and Strehl of 0.977"
More importantly, I get excellent results from my 18" Obsession in the field
at dark skies--several in this group have looked through my scope and shared
this opinion. For VISUAL, I'll put it up against any refractor,or Mak
variant, costing with mount under $20,000. Photography is, of course,
another matter. There I'd want a big first rate refractor--an AP, a TMB or
Tak. Marcus's comment, btw, was that "the interferometer specs must be fake
or measured improperly"--BS--I think I know who the fake is.
Mike Barrs <mba...@NOSPAM.mangonet.com> wrote in message
news:05Tq4.9476$QK6.1...@news4.mia...
JMc
why, I wonder, would they want to exclude data about the edge zone of
large mirrors? Do their customers just mask off the outer 1/2 inch?
Jim McSheehy
Andy Wallace
as accurate if you are foolish enough to believe everything he says--again,
I would invite you to look through my scope.
JMc <tri...@my-deja.com> wrote in message
news:38AC17B8...@my-deja.com...
Ratboy99
>every Joe to brag about his "1/29.3" wave mirror ?
To sell more of his fine mirrors, of course.
Roland Christen DOES
>have all the numbers, and real ones too (off the interferometer) for his
>APOs and he does indeed keep them for himself only.
Yep. You're right.
>I really think Zambuto and John Hall are shooting themselves in the foot
>by supplying those (largely imaginary) numbers with mirrors they make.
>
>Bratislav
Yeah right. Looks like they're crying all the way to the bank!
rat
~( );>
WHALEN44
And he got it from their web page, which states the following:
" All optics manufactured by GALAXY OPTICS are guaranteed to be diffraction
limited (including obstruction) with a maximum RMS wavefront error of 0.060 for
18.0" optics and 0.070 for the larger sizes, errors are measured at 632 nm over
a 98% plus clear aperture."
That being said, I for one still think Galaxy Optics are some of the best
available to the amateur astronomer. I have looked through too many excellent
dobs that used their optics that really "wowed" me to say otherwise.
Richard Whalen
whal...@aol.com
"Time spent observing the heavens is not deducted from your lifespan"
Ratboy99
>>by supplying those (largely imaginary) numbers with mirrors they make.
>>
>>Bratislav
The thing is, Brat, that people like Carl Zambuto and John Hall are doing
absolutely great things for the amateur community, by bringing affordable and
high quality optics to the people, and I hope they continue. OTOH, you
"Comrades" put a lot of effort into bringing them down, which IMO is a
disservice. I don't believe that the numbers are, as you said, "largely
imaginary", but that they are the very numbers that are derived from the
testing methods, and are indicative of how mirrors within a sample compare to
each other using the same test methods. How the numbers are interpreted is a
matter of debate, as is clearly stated on the website. The bottom line is
performance. I have heard nothing but excellent things about the empirical
performance qualities of Mr. Zambuto's mirrors. Now tell me, where are better
mirrors for the money available to the amateur community than through these
small producers? If anything, you should be enouraging their efforts. I for
one, after having purchased an excellent 10" F6 Zambuto mirror last year, am
greatly looking forward to receiving my 18" F4.3 this June.
rat
~( );>
Andy Wallace
Assuming that the "less than 2% of clear aperature" referred to by Galaxy is
the mirror's edge, and assuming that, perforce, the mirror must therefore
have a badly turned edge (all large, thin mirrors have some turned edge),
are you aware of ANY web site where a large, thin mirror supplier guarantees
"100% clear aperture coverage, i.e. specifically including the edge" with
his interferometer readout and, if so, pls direct me to that site be it US,
Russian, German or whatever.
BTW, as you know, masking perhaps 1/4" of the outer edge of a large thin
mirror is not going to have an effect detectible to the naked eye on mirror
brightness but I have not been able to detect any visible need to mask my
mirror. And, if I could save $4,000-$8,000 with this masking process, it
would seem to give me a pretty good return on my hour's work, don't you
think? :)
Andy
WHALEN44 <whal...@aol.com> wrote in message
news:20000217114205...@ng-cb1.aol.com...
Tymewayst
apo eater. Doesnt Zambuto make 6" mirrors? If not can they be custom made on
special order?
Jay Reynolds Freeman
> professional telescope optics suppliers, for a 22"-24" Newtonian mirror.
Who'd you ask?
--
Jay Reynolds Freeman -- freeman at netcom dot com -- I speak only for myself.
Valentine Germann
This is a terrific improvement here, including the obstruction
in the evaluation. The total wavefront quality of a Newtonian has
to take into account the primary, secondary and the effects of the
obstruction and spider vanes. I like to think that some of the net
battles I was involved in back in 1995 have led to this result! Some
of you may even remember! In any event, the quality of medium-
and large-sized mirrors have greatly improved since the early 1990s!
As for the edge, Texareau (sp) always masked off the outer 5mm
of his mirrors, stating that the extreme edge could not be quantified.
Val Germann
Central Missouri Astronomical Association
Columbia, Missouri
Mike Barrs
news:88hgo2$qh4$1...@nntp6.atl.mindspring.net...
> > During the past two weeks, I've been getting some quotes from a few
> > professional telescope optics suppliers, for a 22"-24" Newtonian mirror.
>
> Who'd you ask?
So far, I've talked to Torus Technologies (http://www.torusoptics.com/),
Telescope Engineering Company (http://www.telescopengineering.com/), and Bob
Goff at AXE (520-790-1452), who is associated with the Hextek cellular
mirror outfit (http://home.flash.net/~hextek/).
For "amateur" mirror suppliers, I also talked to Steve Swayze, a nice guy,
highly recommended for large custom mirrors
(http://home.europa.com/~swayze/). I tried to reach Ed Beck (Enterprise
Optics) who did the mirror on my 18", but I've had trouble contacting him.
This is turning into a long, involved saga. I'll put up a web page about it
when it's all over.
Mike Barrs
Ratboy99
>
>Let him keep this ONLY for himself, not for us. Why he let us
>know the info which is only for his private use?
>
>
>Valery.
What's the matter Val, forget to take your grouchy pill this week?
rat
~( );>
Enter your name here
through it!!! Frankly my dear, who gives a damn what the test reports
show. If you read the reports in this group, or read Todd or Ed's reviews of
a whole lot of Dobs with mirrors by Carl or John, the response is WOW!!!!
Either we and they are really delusional, or there are too many
anal-retentive types spending too much time worrying about how many angels
are line dancing on the head of that pin to take the time to use the old
Mark I eyeball test. Buy it, enjoy it, find something else to obsess about!
Dave Novoselsky
Enter your name here
Toto on board too) and my 18 should be here in the next few months. Perhaps
we should bring our scopes with their 'crappy optics' out for a shootout
with some of the 'better product?' Hell, come over, we can match them up
with my squadron of 60m thru 152mm Taks, or a few of the other good APOS in
the area! Should be fun. Wonder how much an 18 "professionally built"
mirror would cost, and how many of those would be sold. I guess all of us
how know and like Rick and his products, or own Obsessions, etc. are just a
bunch of deluded fools . . . NAH!!!!!!!!!! Dave Novoselsky
"Ratboy99" <ratb...@aol.comet> wrote in message
news:20000217121800...@ng-cb1.aol.com...
John J. Kasianowicz
And for those who want to read about it, a discussion of the effects of
spider vanes and central obstructions on image contrast was published in a
two-part article (Sky & Telescope; July, 1993 and September, 1993).
Enter your name here
Starmaster EL:
"I won't mince words here: This is an awesome telescope. At Astrofest, I
could not find anything in the field that was significantly better, at any
price level. I do not know what Carl does to these mirrors, but I can say
that every Zambuto-equipped telescope I have ever seen has been a real
stunner. I have trouble tearing myself away from the eyepiece." Of course
Ed appears to have left his techincal gear home with him that day and didn't
measure the mirror. On the other hand, he is hardly a fool or a novice, and
he sure seems to like these 'dime store' optics made in by these "amateurs"
here in the hinterlands. Dave Novoselsky
"Ratboy99" <ratb...@aol.comet> wrote in message
news:20000217175028...@ng-cs1.aol.com...
Ed Erbeck Jr.
You summerised reality! If you plan to use a Scope in Focus. Carl and
Roland deliver what they promise (given atmospheric conditions).
I think I'll bow out of this thread for awhile as I have other matters to
attend to.
--
Ed Erbeck Jr. (Owner Crazy Ed Optical)
Online Catalog:
http://www.crazyedoptical.com
Snail Mail:
Crazy Ed Optical
P.O. Box 110566
Campbell, CA 95011-0566
Phone / Fax 408-364-0944
bratis...@my-deja.com
ratb...@aol.comet (Ratboy99) wrote:
> >So why doesn't he keep 'his' numbers for himself instead giving them
to
> >every Joe to brag about his "1/29.3" wave mirror ?
>
> To sell more of his fine mirrors, of course.
>
> Roland Christen DOES
> >have all the numbers, and real ones too (off the interferometer) for
his
> >APOs and he does indeed keep them for himself only.
>
> Yep. You're right.
>
foot
> >by supplying those (largely imaginary) numbers with mirrors they
make.
> >
> >Bratislav
>
So what about their (also fine) competition from people like Hudek,
Swayze, Spooner, Baldwin, Rokosz etc. who don't feel like "laughing all
the way to the bank" by supplying YES! largely imaginary numbers with
their mirros ? Not even Kodak can compete with someone claiming 1/29.3
mirrors using Pyrex and pitch.
Keep your mirrors, I'm sure they are fine, but I don't really think you
(and few others) understand metrology and need for uniform measurement
methods at all.
Yes, I know, who cares.
Exactly.
rigelsys@home
Me, I'm enjoying the insult fest in poorly worded english among non english
speakers.
All Marcus, Valery et all have done is demonstrate why I would never buy a
mirror
from them, they're collectively south bound ends of a north bound horse!
bye
Leon
ar...@selena.kherson.ua
Dave, Rat,
I am not against quality of Zambuto mirrors. We will see its real
quality soon when it will be tested by interferometer. One mirror
will be tested in Germany by Markus and one I will buy as second
hand and will test it in two labs. We will let know to public these
results.
Concerning impression that 10" mirror of a good quality can bring
to an amateur, I knew far before the s.a.a. and Internet in general
were begun. Keep this nice argument for someone else - he will
listen you with pleasure.
What I can't accept is 1/18.5 , 1/23.6 , 1/26.9 wave front
quality of these mirrors. I simply don't have a permission to publish
here a parts of many letters I received from americans about these
Z. claims. One common thought in all these letters - 95% of
s.a.a. members know well what is going on and only a few anothers
belive in these fantasy figures. This all what I can finally say.
Valery.
valery_...@my-deja.com
"rigelsys@home" <rige...@home.com> wrote:
> Ed,
>
> Me, I'm enjoying the insult fest in poorly worded english among non english
> speakers.
> All Marcus, Valery et all have done is demonstrate why I would never buy a
> mirror
> from them, they're collectively south bound ends of a north bound horse!
>
> bye
>
> Leon
Leon,
FYI, more than twenty 30" mirrors were made by ARIES for the
USA market. They still works just fine, a lot of larger and smaller
mirrors and sytems were made for europian and far-eastern customers.
They all were fully satisfied. You can be 100% sure that you will have
no chance to buy our mirrors - the are simply not for your pocket and
for your tasks and, if you do have them, a large, cheap, thin dob
mirror will serves just fine. We definitely will never come to the market
of a cheap mirrors. This is not our profile. There are a lot guys out
here who making money on them.
If you will be only a bit more attention, you will figure out that I never
said that Z. or any other US-made mirrors are poor. They simply fully
corresponds to a money their customers pay.
The question is in 1/26.9 wave front of these mirrors. If you only
can understand what does it mean 1/26.9 wave, then you will
never write your silly letter.
FYI. 1/26.9 wave front does mean that the mirror does has
smooth deviation from ideal parabola in 0.000009mm .
And these guys proudly claims that they can mechanically mesure
it. A Fantasy land in the reality.
teletex
Please tell me how to order one of your 12.5 F5 mirrors, I
am recieving a 12.5 Portaball (Zambuto Optics) and would
like to compare your optic to his. I am an experenced observer
and will tell the absolute truth about what I see.
I will also have other observers over for the evaluation.
I know you make great optics and think the comparison would
be fun and informing. If you would make me a mirror please tell
me how to order it. It would be a under the stars test where
it really counts!
Thanks,
Bob
* Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
The fastest and easiest way to search and participate in Usenet - Free!
David A. Novoselsky
the belt. Valery and Markus' English is a damn sight better than my tortured
German or my almost nonexistent Russian (a few word, mostly profane.) I suspect
you are not as well versed in either of their native tongues as they are in your
language. They manage to communicate in a tongue that is foreign to them and on
a subject that is arcane and highly-technical a whole lot better than many
"natives" on this newsgroup. I may not agree with what they say, but taking a
shot at them for their use or awkward misuse of the English language is not
funny, it is just Beyond the Pale. I think it is time to call the Thread Police
and cut this particular form of nonsense off before it grows into a cancer that
stifles this group -- besides, watching the fur fly between Markus and Valery,
and occasionally Roland, Ratboy, and others, is a whole lot more fun than
watching TV and the seeing is too lousy to go outside. Let's keep it above the
belt, okay? Dave
ar...@selena.kherson.ua
teletex <karlindse...@earthlink.net.invalid> wrote:
> Valery,
>
> Please tell me how to order one of your 12.5 F5 mirrors, I
> am recieving a 12.5 Portaball (Zambuto Optics) and would
> like to compare your optic to his. I am an experenced observer
> and will tell the absolute truth about what I see.
> I will also have other observers over for the evaluation.
>
> I know you make great optics and think the comparison would
> be fun and informing. If you would make me a mirror please tell
> me how to order it. It would be a under the stars test where
> it really counts!
>
> Thanks,
> Bob
Bob,
Some points:
1.We need to know your requiements - then we can give you our
price.
2.If this will be only one mirror - this will be even much more
expensive.
3. The best way to keep your money for something another more useful,
then simple ARIES mirror verification, simply give your mirror to
a test optical lab and see the interferogram. I will send you per
advance some our typical interferogram for a 16" F/3 mirrors
(at least 4 times more difficult to make).
4. If the precision you would like to have is better than 1/12 wave
front, I would advice you to order 10" or even 12" Mak-cass from
Roland Christen.
The price will be very similar (because of custom order for the
mirror).
5. If you need a simple amateur level mirror, you can ask Intes.
This advice I always gave to Markus when he try to order from us
1/4 - 1/8 wave 12" mirrors.
Take care.
ar...@selena.kherson.ua
"Mike Barrs" <mba...@NOSPAM.mangonet.com> wrote:
> <bratis...@my-deja.com> wrote in message
> news:88fv38$e56$1...@nnrp1.deja.com...
> > In article <20000216005532...@ng-fl1.aol.com>,
> > ratb...@aol.comet (Ratboy99) wrote:
> > > >I get the sense that Zambuto uses his test to perfect his
mirrors,
> > and
> > > >discourages folks from reading too much into them.
> > >
> > > Excellent point. I think the numbers give Carl a consistent
benchmark
> > to use
> > > for figuring the mirrors. They are more for his use than for ours.
> >
to
DOES
> > have all the numbers, and real ones too (off the interferometer) for
his
> > APOs and he does indeed keep them for himself only.
foot
> > by supplying those (largely imaginary) numbers with mirrors they
make.
> >
> > Bratislav
>
wavefront, and
> RMS wavefront. One set is called "indicated error, uncorrected for the
> effect of random variables", and the other set is called "corrected to
> simulate potential effect of random variables". That's an interesting
> marketing approach. It gives you a choice of being realistic when
talking
> about your optics, or else using inflated numbers for bragging rights.
:-)
> One recent large Pegasus mirror I saw had 1/8 wavefront P-V as the
corrected
> number, and 1/13 wavefront P-V as the uncorrected number.
>
> professional telescope optics suppliers, for a 22"-24" Newtonian
mirror.
the
> typical large amateur mirror from Pegasus, Nova, or Galaxy. It's a
real
> eye-opener. Quotes for figuring 1/8 - 1/10 wavefront are running from
$6,000
> to $13,000, and that's not including the cost of the blank! It does
make one
> just a wee bit suspicious about the wavefront claims for amateur
mirrors
> costing 1/3 to 1/4 as much.
>
> Why is it so expensive to get a mirror from these pro optical shops?
Well,
> for one thing, their clients are mostly government agencies and
> universities. These outfits either have their own testing instruments
to
> confirm quality, or else they subcontract for independent testing of
the
> delivered product. That keeps the wavefront numbers honest.
Independent
> testing and confirmation is the missing ingredient in the amateur
mirror
> market.
>
most of
> the time. But this test report numbers game is confusing, and in some
cases
> it borders on being dishonest (I know of two recent mirrors from a
highly
> reputable amateur supplier that failed a star test, by a wide margin,
and
> both mirrors had excellent test report numbers).
>
> An outfit like Zambuto, Pegasus, Galaxy, or Nova should simply give a
> money-back satisfaction guarantee that a new mirror will pass a star
test,
> period. Test report numbers are worthless, unless we have the ability
to
> duplicate and confirm the numbers. The star test is the amateur's
equivalent
> of a Zygo interferometer. It's the only realistic certification
procedure we
> have.
>
> Mike Barrs
Mike,
I agree with all points in your post. But I can't consider a StarTest
as Zygo interferometer amateur's equivalent. Much better for amateurs
will be another step. If they like to be sure in their mirrors quality
they can spend $200-400 for independent interferometrical test. And
if their mirrors will not pass, say, 1/6 wave front they must ask their
money back + test cost. Such step will keep most these manufacturers
far from such brave claims like they doing.
Take care,
RAnder3127
>mirrors using Pyrex and pitch.
>Keep your mirrors, I'm sure they are fine, but I don't really think you
>(and few others) understand metrology and need for uniform measurement
>methods at all.
>Yes, I know, who cares.
Let's say the mfg. actually manages
to fabricate a mirror to that level of
perfection. Then he takes it out of the
lab. Does it keep to that level
when it's mounted in one of those
el-cheapo Dob mirror mounting
apparatus, like a cloth sling? Is the large, thin mirror's shape
maintained to 1/29.3 wave when the scope
is re-positioned and it's tilt changed?
Not likely.
-Rich
Chris1011
level of performance these test out, so that I might be enlightened as to what
specific wavefront error can be had and still provide the mysterious and
elusive 'diffraction limited' performance that>>
I was curious to see how interferometry compares to the focault test, so i did
some testing last night. (by the way, sorry for being testy myself in the last
few days - I had a wisdom tooth pulled and apparently lost 1/4 of my wisdom in
the process).
I have a bunch of 8" F4.5 mirrors that were finished to between 1/10 and 1/12
wave P-V as measured on the interferometer. I tested them with a knife edge
with a zonal mask and came out with vastly smaller errors, on the order of 1/30
wave P-V. I looked up in Malakara again to see what was the difference in the
test methods, and I think I know what is happening.
With the Focault test, a bunch of zones of nonzero dimension are read in one
dimension (either horizontally or vertically) on the mirror surface. These are
discrete areas on the mirror whose average radius is measured. The entire
mirror is not measured, only a limited number of areas.
With the interferometer, the entire surface is measured, at as many as 300
points. In fact, some interferometers automatically measure tens of thousands
of points to get an accurate profile of even the most minute area. Furthermore,
the interferometer measures astigmatism and residual coma, which the Focault
test does not do, at least not without an ungodly amount of work.
As an example, I took one mirror which has the following interferometrical
data: Astigmatism=.056, Coma=.039, Spherical=.042, all in waves P-V. Overall
P-V=.080, or about 1/12 P-V.
With the Focault test, only the spherical component would be measured, so the
mirror would come out 1/23 wave P-V. Actually, at this focal ratio, I could
not read the knife edge accurately enough to be sure of this level of accuracy.
In fact, it looked better than 1/23 with the knife edge.
My conclusion is that the interferometer values are somewhat overly pessimistic
and the knife edge is somewhat overly optimistic on mirror quality, and that
half way in between for each is probably closer to the truth. I would say that
a 1/30 wave Focault tested mirror will probably test on the interferometer
between 1/10 and 1/14 P-V., a not too shabby mirror. A close reading of
Malakara shows that this is the case. He does not favor the use of P-V at all,
and says that RMS readings are more representative of actual performance. A
mirror with 1/14 wave RMS would be just barely acceptable, and 1/50 RMS would
be pretty much perfect.
Roland Christen
ASTRO-PHYSICS
Andy Wallace
turned down edge and, if so, how is that edge performance quantified?
Thanks, Andy
Chris1011 <chri...@aol.com> wrote in message
news:20000219103520...@ng-fb1.aol.com...
Jan Owen
Here, for the first time, we see something that begins to make some sense
out of all the seemingly conflicting data!
----------------------------------------------------------------------------
-------------------------
Chris1011 wrote in message <20000219103520...@ng-fb1.aol.com>...
>With the interferometer, the entire surface is measured, at as many as 300
>points. In fact, some interferometers automatically measure tens of
thousands
>of points to get an accurate profile of even the most minute area.
Furthermore,
>the interferometer measures astigmatism and residual coma, which the
Focault
>test does not do, at least not without an ungodly amount of work.
>
Ron Wodaski
should always be tempered by someone actually going out and trying to do it
to see what, in fact, happens. <g>
Ron
"Jan Owen" <Jan....@gte.net> wrote in message
news:7xzr4.627$TM2....@dfiatx1-snr1.gtei.net...
rob hunt
>The reason I put the word "CRAPPY" in single
>quotes was ...
>simply my own hyperbole -- intended to overstate to make a
point.
hmmm... 'hyperbole -- intended to overstate to make a point.'
it seems to me that's the WHOLE problem. zambuto GROSSLY
overstates his performance figures and manufacturers who don't
are left with several unpleasant choices.
should they publish their own objective test data and have it
look pale next to someone else's exagerrated claims?
maybe they should say nothing at all and just let the chips fall
where they may.
the third choice of trying to point out the situation to
the 'community' is obviously an uphill battle in a world that
puts so much faith in the occult.
i am not a scientist and at least for me this hobby is mostly an
aesthetic pursuit but it is science and engineering that give us
the tools we use and those fields have certain rules. uniform
standards are part of it and i can't for the life of me
understand why people in this newsgroup would be seeking
anything less.
no-one has ever said zambuto doesn't make good or even excellent
mirrors so i think we can leave the word 'crappy' all the way
out of it. i've seen almost a dozen of his larger mirrors in
action and i was impressed by the view every time. none of that
is even the point. the point is are you responsible with
numbers or do you treat them like the US congress?
robert hunt
www.goes.com/~robhunt
Bryan Greer
> My conclusion is that the interferometer values are somewhat overly
> pessimistic and the knife edge is somewhat overly optimistic on mirror
> quality, and that half way in between for each is probably closer to the
> truth.
I understand your point about how the Foucault test misses four of the five Seidel
aberrations. The undersampling of the Foucault test, along with operator
variability, are the two main sources of error. However, I don't quite see why you
came to the conclusion that an interferometric test is overly "pessimistic". Are
you saying this is a general truism, or just in this particular case? My
experience with interferometric testing is that it represents something closer to
reality. If we assume the test is set up properly, I don't see why it would
systematically yield numbers worse than reality.
I can provide another anecdotal example similar to yours. A 10" f/4.5 mirror was
produced locally by an experienced mirror maker, Bill Herbert. The mirror was one
of a batch of six, and this one turned out the best (a little better than 1/20th
P-V per Foucault). Dick Suiter has this mirror now, and it was the one used in his
recent ATMJ article describing a video camera implementation of the Foucault test.
To validate the video technique, he had Bill Zmek of Perkins-Elmer
interferometrically test the mirror with their Zygo. The Zygo also read a little
over 1/20th P-V. I don't mention this to dispute that the Foucault usually gives
an overly optimistic result--I completely concur with you on that. In this case it
wasn't true, but that's probably just because Bill and Dick are very good operators
at the knife-edge (or maybe blind luck!). I'm just not sure I agree that it's
generally true that the interferometer yields overly pessimistic numbers.
Thanks for the test results. We need more of this on s.a.a.
(BTW, my above comments should not be interpreted as a criticism of the Foucault
test. It is the best in-process test for making a paraboloid.)
Sincerely,
Bryan Greer
Columbus, OH
(please remove *takeout* from my address for private replies)
Zane
(snip)
> The
>resulting interferograms measure the accuracy of the ENTIRE TEST CAVITY,
>including the autocollimating flat, diverger optics, beamsplitter, and
>reference arm optics. Since errors in any one of these areas will tend to make
>the overall test result worse (negative correlations
>with residual errors in the optic under test being more probable than positive
>correlations), it is fair to conclude that an interferogram establishes THE
>WORST possible figure accuracy of the optic under test. There is thus a very
>real possibility that the actual figure accuracy of
>your mirror is BETTER than the test results indicate."
I noticed someone posted that they were going to do some interferometric
tests on a couple of Zambuto mirrors. I think one of the parties also
talked about how cheap an interferometer is. Everyone should be prepared
to take the results with several grains of salt, I think, for the above
reasons and others. You don't just do a quick test and get accurate
numbers with a "cheap" interferometer. Not to mention the environmental
control, etc.
Zane
John J. Kasianowicz
other form of harm. I think he's being sincere here.
Andy Wallace
part of anyone--it has been about comparing apples to oranges to lemons.
Zane <zane...@sansnetcom.com> wrote in message
news:38aed1b3...@nntp.ix.netcom.com...
> chri...@aol.com (Chris1011) wrote:
>
> (snip)
>
> > A close reading of
> >Malakara shows that this is the case. He does not favor the use of P-V at
all,
> >and says that RMS readings are more representative of actual performance.
A
> >mirror with 1/14 wave RMS would be just barely acceptable, and 1/50 RMS
would
> >be pretty much perfect.
>
> I was wondering when someone would bring this up. Using just a P-V number
> requires an assumption to be made about what the type of error is and
> leaves out some things that can be important in image quality.
>
> Zane
Chris1011
I understand your point about how the Foucault test misses four of the five
Seidel
aberrations. The undersampling of the Foucault test, along with operator
variability, are the two main sources of error. However, I don't quite see why
you
came to the conclusion that an interferometric test is overly "pessimistic".
Are
you saying this is a general truism, or just in this particular case? >>
In the case of Perkin-Elmer, I would assume that they have a very expensive
system costing in excess of $1 mil, which does phase measurement and can be
accurate to perhaps 1/100 wave. This is not the case with my "cheap" $50K
system. I have characterized my interferometer's errors to be on the order of
1/25 wave. However, it is very useful in showing where the corrections need to
be made, and shows astigmatism very easily. I use it as a quality control
device, not as an absolute measurement. When the optic under test is better
than 1/10 wave including the interferometer errors, I can stop messing with it
and get it ready for sale.
In the case of interferometer P-V measurements, every little departure from
perfection adds to the error, even if it has little or no effect on the final
image. The RMS measurement takes this into account, and reduces the effect of
small localized errors on the final number. Unfortunately, this is not easily
computed with the Foucault test.
Roland Christen
ASTRO-PHYSICS
Chris1011
Does your interferometer read 100% of the mirror--i.e., does it test for a
turned down edge and, if so, how is that edge performance quantified?
>>
Yes it does measure right to the edge. However, if the CCD optics are not
perfectly focused, some interference effects will result which can make perfect
edge measurements somewhat less than accurate. The edge is better judged by
autocollimation with the star test. you can see right away a 1/8 wave turned
edge as an increase in the fuzz level.
Roland Christen
David A. Novoselsky
even more of a pain in the a-- than we normally may be. A healthy sense of
irreverence wouldn't hurt any of us. Smile at the easy part of life, worry
about the really important issues. Dave
Andy Wallace wrote:
> John--
>
> I was kidding Dave--all lawyers are required to buy their first NOMEX suits
> when they graduate law school. :-) And lawyers probably use hyperbole more
> than any other group--it can be very effective :-)
>
> Andy
>
> John J. Kasianowicz <sur...@erols.com> wrote in message
> news:88mm0m$n04$1...@bob.news.rcn.net...
lude...@my-deja.com
Hi Roland,
a great idea to do such comparation, so my Portaball have an real
chance to use a real ~ 1/8 wave mirror, which is more than I expected
during placing my order.
great Post, thanks
Markus
> My conclusion is that the interferometer values are somewhat overly
pessimistic
> and the knife edge is somewhat overly optimistic on mirror quality,
and that
> half way in between for each is probably closer to the truth. I would
say that
> a 1/30 wave Focault tested mirror will probably test on the
interferometer
> between 1/10 and 1/14 P-V., a not too shabby mirror. A close reading
of
> Malakara shows that this is the case. He does not favor the use of P-
V at all,
> and says that RMS readings are more representative of actual
performance. A
> mirror with 1/14 wave RMS would be just barely acceptable, and 1/50
RMS would
> be pretty much perfect.
>
> Roland Christen
> ASTRO-PHYSICS
Mike Barrs
> test, so i did some testing last night.
<test results snipped>
> My conclusion is that the interferometer values are somewhat overly
> pessimistic and the knife edge is somewhat overly optimistic on mirror
> quality, and that half way in between for each is probably closer to the
> truth.
I'd like to add my thanks here, for running that comparison. Your conclusion
helps clear up some things I was wondering about.
As I mentioned earlier in this thread, I've been getting some quotes from
some of the higher-end optical shops, for a 22", F/4 mirror (2" thickness).
The quote I got from Torus Technologies was interesting. They said they
could *only* produce an interferometer-certified mirror in lambda/8 surface
quality (1/4 wavefront) in that size and thickness. It's not a limitation in
their figuring or testing capabilities, but a limitation in the stiffness of
the mirror material. They have the capability of doing 1/10 wavefront
optics, but they only do this on thick optical flats (for high powered laser
applications). Thermal effects might also have something to do with it...
but I asked about Zerodur instead of Pyrex, and they said we'd have the same
problem pushing past a lambda/8 surface in Zerodur.
That surprised me, especially when we're talking about a $12,000 to $16,000
mirror here (depending on whether it's Pyrex or Zerodur). At those prices,
you expect high optical quality.
In the amateur astronomy community, you'll hear many recommendations from
the mirror suppliers and telescope builders to get at least 1/8 wave quality
in your mirror. Optics of "only" 1/4 wave quality are gently disdained. For
example, page 199 in Suiter's star testing book... "For most people who
observer under average skies, a 1/4 wavelength correction error represents
an acceptable compromise between quality and the price of optics." And
again on the same page... "Personally, I find images of optics that are
nudged against the Rayleigh limit a bit too soft". In other words, you need
1/8 wave optics or better, otherwise you have a crappy telescope. Read the
Starsplitter, Starmaster, or Pegasus web pages, and you'll see basically the
same thing. How does that reconcile with the quote from Torus for a $12,000
1/4 wave mirror?
Roland's conclusion from the comparison test helps put this in perspective.
If I understand this correctly, an interferometer-certified mirror at 1/4
wavefront is roughly equivalent to what most amateurs are calling a 1/8
wavefront mirror (in terms of true performance and star testing), and also
roughly equivalent to what suppliers of amateur mirrors call a 1/8 to 1/12
wavefront mirror, based on Focault testing.
That makes sense to me (I think). Considering the physical specs (22"
diameter, F/4, 2" thickness), a 1/4 wavefront rating by interferometer for
this mirror probably isn't too shabby. ;-)
> Furthermore, the interferometer measures astigmatism and
> residual coma, which the Focault test does not do, at least not
> without an ungodly amount of work.
This is also a significant comment. Pay attention, folks. I recently saw a
mirror from a reputable mirror supplier that had objectionable astigmatism,
even though it came with a Focault test report claiming a high wave rating.
I strongly recommend doing a star test for astigmatism after receiving a new
mirror.
Mike Barrs
Andy Wallace
Interesting. I have an interferogram on my 18" mirror that shows 1/6.5 P-V
wavefront over in excess of 98% of the mirror--measured at 2900 points with
a Strehl of 0.977. Assuming, as has been suggested, the less than 2% is the
mirror edge and that it is turned, by masking perhaps the outer quarter, do
I have a 1/13 wavefront in Focault?
One thing I have learned in 35 years of law practice is to take so-called
"expert" testimony with a very large grain of salt--it is frequently all
over the map on the same subject. I think we do get right back to the best
startest acceptable to the scope owner for the most reasonable amount of
money. This current "spec" war thus seems a little silly. As for the
upcoming "face-offs", Todd Gross has already done this. The 18" dob won :-).
Andy
Mike Barrs <mba...@NOSPAM.mangonet.com> wrote in message
news:dHWr4.734$Wv2....@news2.mia...
> That makes sense to me (I think). Considering the physical specs (22"
> diameter, F/4, 2" thickness), a 1/4 wavefront rating by interferometer for
> this mirror probably isn't too shabby. ;-)
> This is also a significant comment. Pay attention, folks. I recently saw a
Andy Wallace
Andy Wallace <andyw...@home.com> wrote in message
news:GkXr4.1588$gE6.1...@news1.rdc2.tx.home.com...
William R. Meyers
Hi, Mike,
Thanks for an exceptionally lucid and helpful post. I learned a lot from it.
I only wish I were in the market for a $16,000 mirror of 22" diameter! But
your suggestions apply to smaller mirrors also.
Cordially,
Bill Meyers
Mike Barrs wrote:
> That makes sense to me (I think). Considering the physical specs (22"
> diameter, F/4, 2" thickness), a 1/4 wavefront rating by interferometer for
> this mirror probably isn't too shabby. ;-)
>
> > Furthermore, the interferometer measures astigmatism and
> > residual coma, which the Focault test does not do, at least not
> > without an ungodly amount of work.
>
randy_...@my-deja.com
numbers used for in-process figuring be releashed
as (to be misunderstood to be) real peak to valley
values for the entire wavefront. The zonal test
maybe reliable and repeatable but it doesn't
represent the true wavefront pv. I use a phase
measuring interferometer in my fabrication and it
samples 10000 points of the wavefront - it always
gives a higher pv than, for example, an
interferometric fringe analysis. This phase
measurement is as close as it gets to reality
(keeping in mind all the problems you have to deal
with in interferometry). It will deal with the
edge hook up to the mirror diameter divided by
256. That's an example - you can't expect 100K
interferometers to be used for this purpose. I
don't see an easy solution to this issue. The fact
is that the mirrors are undoubtably excellent but
what are they really? (if you have to throw
numbers around) BTW, I could offer to test a
Zambuto mirror (in fact I do) but that would be
just a single data point and many mirrors would
have to be tested for any correlation to be
apparent. And, of course, there is a mirror maker
bias against interferometry (for several valid
reasons but also some fear of what may be found I
think). For the users (but not the pundits!) the
eye is the ultimate interferometer.
Randy
> > ~( );>
randy_...@my-deja.com
wondered about these incredible numbers. I run an (US Navy) optical
fabrication facility. My primary optical metrology is phase
interferometry. In addition to the interferometer itself one must have
accurate reference surfaces and optics (such as compensators) along
with good stages and linear slides. This is expensive stuff, but
obviously not necessary for skilled persons to make good mirrors. In the
industry (the real optics industry) peak to valley is associated with
the entire surface being tested in some direct and fairly complete
fashion (testplate, fringe interferometer, phase interferometer).
Aspheric optics are commonly tested with a profilometer - this is a test
that covers just a few "scans" of the surface and generally yields an
optimistic value. And, it pretty much goes without saying that
knife-edge results are strictly "inprocess data" and buyoff data is
usually a test that involves interferometric wavefront analysis, MTF,
encircled energy or maybe all of them. So, I feel that you simply can
not back a wavefront SURFACE peak to valley value out of a test that
samples only a small faction of the surface area, and then the first
derivative of the surface at that. And out of focus, etc. At the least
it has to be statistically qualified in terms of the area covered to
make any metrological sense.
I have a couple of stories too. Back in the early 80's I watched the
painful transition from test-plating to computer fringe analysis.
Opticians had to adjust as their 1/4 fringe parts were being rejected at
0.135 waves ... or 0.2 waves! Then in the late 80's phase
interferometers repeated this process. There are still some crusty
opticians around that believe there is something fundamentally wrong
here.
I have interferometerically tested several amateur made mirrors and
have faced the dismay of the owner when I utter something like 0.22 pv
and he was expecting 0.05! Though after seeing the problems rotate with
the mirror they usually agree. Every time the correction (SA3) has been
very good.
Finally, I have to make an observation. Until recently I can't remember
too many outfits making (direct numerical) wavefront claims without the
means of directly measuring it. Edmund used to claim some surface
accuracy way back before they had an interferometer. Anyway, outfits
usually claimed something like diffraction limited/theorectical limited
- something without numbers. It was as if there was the common wisdom to
not read too much into the zonal test results. Or maybe it allowed them
to sell lousy optics without too much guilt. In that sense I'd rather
see the situation we have now where these numbers indicate high quality
optics are being made.
Best,
Randy
In article <CrXr4.1595$gE6.1...@news1.rdc2.tx.home.com>,
Mike Spooner
<randy_...@my-deja.com> wrote in message
news:88qle0$gu4$1...@nnrp1.deja.com...
> apparent. And, of course, there is a mirror maker
> bias against interferometry (for several valid
> reasons but also some fear of what may be found I
> think). For the users (but not the pundits!) the
> eye is the ultimate interferometer.
> Randy
>
Hi Randy, Thanks for some good info in your posts (along with many others).
I'm not sure there is all that much bias against interferometry by opticians
perhaps, as much as it has been ingrained that 1/10 wave is magic number and
nothing less is worth bothering with. Even Suiter was quoted here as writing
(and he did) about 1/4 wavefront optics being close to suspect but I believe
he was dealing with a full aperture case of relatively smooth correction
error (low order SA). This is quite different than a 1/4 wave (or 1/2 wave
error) that only occupied a small fraction of the area. Others here have
mentioned RMS as a preferred measure and I quite agree.
--Mike Spooner
Sue_and_Alan_French
People should read "Optical Quality in Telescopes" in the March, 1992 issue
of Sky & Telescope, and pay close attention to what Douglas George and
Terence Dickinson say about the differences between a true 1/4 wave and true
1/10 wave telescope.
Clear skies, Alan
Mike Spooner wrote in message ...
>
[SNIP]
>I'm not sure there is all that much bias against interferometry by
opticians
>perhaps, as much as it has been ingrained that 1/10 wave is magic number
and
>nothing less is worth bothering with. [SNIP]
Bryan Greer
> So, I feel that you simply can not back a wavefront
> SURFACE peak to valley value out of a test that samples
> only a small faction of the surface area, and then the first
> derivative of the surface at that. And out of focus, etc. At
> the least it has to be statistically qualified in terms of the
> area covered to make any metrological sense.
>
What you are saying is so true, Randy, and it has an analogous
counterpart when describing microroughness. In Jean Bennett's "Surface
Roughness and Scattering", she says (pg. 38):
"…the rms value will depend on the size of the areas. For these
reasons, (italics/bold) there is no unique rms roughness value for a
surface (end italics/bold). It depends on the following:
1. The length of the surface profile (maximum spatial wavelength).
2. The surface area being averaged over for each measurement
(lateral resolution), and
3. The distance between data points (sampling distance)."
My own measurements of optical surfaces taken with the atomic force
microscope bare this out. The same surface can have significantly
different rms roughness values depending on the size of the area
sampled. The methods we chose to quantify a optical surface have
everything to do with the numbers we get. This isn't a criticism of any
particular test. Rather, as you say, it simply points out the need to
specify what method was used, and an understanding of what is a
reasonable level of precision to expect. The latter is not always so
straightforward. I've yet to run across a really complete analysis of
the Foucault test (using Taguchi's statistical methods or similar).
> I have interferometerically tested several amateur made
> mirrors and have faced the dismay of the owner when I
> utter something like 0.22 pv and he was expecting 0.05!
> Though after seeing the problems rotate with the mirror
> they usually agree. Every time the correction (SA3) has
> been very good.
>
>
Experienced star testers know this. It's so rare to see an optical
system that even approaches some of the accuracy numbers tossed around
so freely by some. I can live the rest of my life with a telescope that
regularly delivers an honest 1/6 S.A.
> In that sense I'd rather see the situation we have now
> where these numbers indicate high quality optics are
> being made.
>
That is a very good point, and one I was discussing on the phone with a
friend the other day. Don't we all remember the "dark ages" (1970s and
early 80s) of commercial telescope optics? For the most part, the
commercial mirrors made today by the various small enterprise shops are
vastly better than what we got back then. I'm not sure the amateur
community really appreciates the bargain prices they're paying for top
grade optics. We can quibble about their testing methods, etc., but not
many can (or are) disputing their quality!
> To get back to the subject: Why let internal PV
> numbers used for in-process figuring be releashed
> as (to be misunderstood to be) real peak to valley
> values for the entire wavefront.
>
Personally, I think it's great that some of these guys are giving us the
Foucault numbers. It just has to be accompanied by enough "educational
material" so it's not "misunderstood", as you say. When I make a
mirror, and the Foucault says 1/30, I'm happy. I don't believe my
mirror is really 1/30, but I know it's mostly likely diffraction
limited, and will not be the bottleneck in the optical system.
> This phase measurement is as close as it gets
> to reality
>
I pointed out this same thing about a year ago when Markus and Roland
were running into a discrepancy between interferometric tests. A full
Fourier analysis of the fringes is a leap closer to reality compared to
a "fringe center" type analysis. Again, both tests have their place and
purpose, so this is not a criticism. It's just to point out that all
interferometers are not the same.
The Beast
"Sue_and_Alan_French" <Sue_and_A...@email.msn.com> wrote:
> Hi Mike,
>
> People should read "Optical Quality in Telescopes" in the March, 1992
issue
> of Sky & Telescope, and pay close attention to what Douglas George and
> Terence Dickinson say about the differences between a true 1/4 wave
and true
> 1/10 wave telescope.
>
> Clear skies, Alan
>
>
glass? No offense intended, but I'm sure he is aware of any differences
that would be mentioned in any article in a magazine as non-technical
as S&T.
-----------------------------------------------------------------------
The Beast Appears In Darkest Nights,
Astronomers Gather At Their Sites,
And When They Trod Upon The Grass,
The Beast Will Come To Kick Their Scopes and Eyepieces and Stuff.
Andy Wallace
Sue_and_Alan_French <Sue_and_A...@email.msn.com> wrote in message
news:u9AVUmHf$GA.191@cpmsnbbsa03...
> Hi Mike,
>
> People should read "Optical Quality in Telescopes" in the March, 1992
issue
> of Sky & Telescope, and pay close attention to what Douglas George and
> Terence Dickinson say about the differences between a true 1/4 wave and
true
> 1/10 wave telescope.
>
> Clear skies, Alan
>
randy_...@my-deja.com
I've had the pleasure (rigorous, to be sure) to have worked with Dr.
Jean Bennett on and off over the last 10 years. She is now retired but
still maintains her lab here in the Research Department at China Lake.
There is room for more applied theory here - at least in presenting it
in a clear and understandable way. For all I know she may have analyzed
this very subject. I'll try to talk to her soon about it.
Randy
In article <38b14f3f@news>,
Sue_and_Alan_French
I really think people should dig up the issue and read the entire piece -
most large libraries have Sky & Tel and many college libraries have them
back for many years.
The basic idea was to compare telescopes of known quality. Peter Ceravolo
made 6" f/8 Newtonians with primary mirrors that were near 1 wave, 1/2 wave,
1/4 wave, and 1/10 wave PTV at the wavefront. They were tested by
interferometer. Douglas George and Terence Dickinson were given the scopes
to evaluate. They did not know which was which initially, but were able to
rank them through their observations.
Here are a couple of pertinent passages...
"Clearly this exercise suggest that there are huge gains in visual
performance as optical quality improves to 1/4 wavefront, and less
conspicuous benefits with further improvement."
Terrence Dickinson
"...The 1/4 and 1/10 wavefront mirrors in A and D, restively, only
showed their differences with careful scrutiny under excellent seeing. The
fact that I could tell the difference between them, however, suggests than
an absolute 'diffraction limit' is indeed a myth."
Douglas George
Again, I highly recommend that people make the effort to find the complete
article. I think it is must reading for anyone interested in telescope
performance.
Clear skies, Alan
Andy Wallace wrote in message ...
WHALEN44
Good to hear from you. Your imput here is always welcome and informative. It is
nice to hear from a real pro the skinny on testing etc. Hope all is well with
you and yours.
Richard Whalen
whal...@aol.com
"Time spent observing the heavens is not deducted from your lifespan"
Andy Wallace
Thanks. A curious choice of mirrors in the comparison--what happened to 1/6
wave or 1/8 wave (the subjects, mainly, of the current debate)? Presumably,
it would be even more difficult to differentiate between a 1/6 or 1/8 and
the 1/10 benchmark.
Andy
Sue_and_Alan_French <Sue_and_A...@email.msn.com> wrote in message
news:#Jk#egIf$GA.264@cpmsnbbsa04...
TMBack
>In Jean Bennett's "Surface Roughness and Scattering",
>she says (pg. 38): "…the rms value will depend on the
>size of the areas. For these reasons, (italics/bold)
>there is no unique rms roughness value for a surface
>(end italics/bold). It depends on the following:
>
>1. The length of the surface profile (maximum spatial
>wavelength).
>2. The surface area being averaged over for each
measurement (lateral resolution), and
>3. The distance between data points (sampling
>distance)."
Where in the world do you find books like this Bryan?
You don't buy them, like I do, do you?
Also, on page 39, Ms. Bennett states:
"The same surface can have many different values for
the rms wavefront roughness depending on the instruments
used to make the measurements."
These are surface roughness measurements (resolution
down on the order of 1A RMS), and while I would love
to have a Newtonian mirror with such smoothness, we are
well away from the day that amateur mirrors are this
smooth.
>The methods we chose to quantify a optical surface
>have everything to do with the numbers we get.
Point well taken. However, these threads about optical
wavefront quality are starting to give the impression that
the interferometer is a flawed tool.
>Experienced star testers know this. It's so rare to see
>an optical system that even approaches some of the
>accuracy numbers tossed around so freely by some.
Agreed.
>I can live the rest of my life with a telescope that
>regularly delivers an honest 1/6 S.A.
S.A.? Wavefront yes, but S.A.?
>When I make a mirror, and the Foucault says 1/30, I'm
>happy. I don't believe my mirror is really 1/30
I hope not! But you might want to tell this to your
Columbus group. I had for too many years tried to
explain to some of them that their Admir numbers
were, well, a bit optimistic.
Thomas Back
TMB Optical
Sue_and_Alan_French
Yes, I know of Mike's reputation.
Please note I started my note with "People should..." I was just adding to
Mike's comments, not suggesting he needed to read it.
Clear skies, Alan
The Beast wrote in message <88rt57$cfd$1...@nnrp1.deja.com>...
>In article <u9AVUmHf$GA.191@cpmsnbbsa03>,
> "Sue_and_Alan_French" <Sue_and_A...@email.msn.com> wrote:
>> Hi Mike,
>>
>> People should read "Optical Quality in Telescopes" in the March, 1992
>issue
>> of Sky & Telescope, and pay close attention to what Douglas George and
>> Terence Dickinson say about the differences between a true 1/4 wave
>and true
>> 1/10 wave telescope.
>>
Sue_and_Alan_French
I thought it was a rather good choice of mirrors. I think cutting it finer
would have been overdoing things considerably.
Actually, I would be interesting in seeing how hard it is to see the
difference between 1/10th at the wavefront and 1/20th at the wavefront. Now
all we have to do is find that 1/20th wavefront mirror <g>.
Clear skies, Alan
Andy Wallace wrote in message ...
Andy Wallace
1/10, it should be virtually impossible, in a properly conducted double
blind test, to distinguish between 1/6 and 1/10. Big difference in cost
between these two mirrors though, particularly in an 18" or 20".
Clear skies, Andy
Sue_and_Alan_French <Sue_and_A...@email.msn.com> wrote in message
news:#0#L$WLf$GA.319@cpmsnbbsa03...
Mike Spooner
for this test purpose. The authors mentioned seeing of 7 on a scale of 10
IIRC which may have been a bit limiting but overall they were impressed by
the views. Of course with Roland's recent efforts at examining Foucault and
interferometric results, it would not be surpising if the true 1/4 wave
scope by Ceravolo performed in the top few percent of telescopes. Perhaps an
interesting offshoot of all this would be to wonder what Lord Raleigh's
theroretical 1/4 wave scope would test on an interferometer! It might shed
some light as to why confusion exists today.
I had a chance to look through the scopes of the S & T test at Stellafane a
couple of years ago but the daylight seeing over the ground was not the best
and the thermal heat problems were at least bad enough to overwhelm the
testing (clouds prematurely ended my attempts). I couldn't stay for the
night due to a tight timetable. :( I would hazard a guess from what I saw
that anyone fortunate enough to get a Ceravolo optic would not be
disappointed. It would be fun to try these scopes in good night skies. (But
then I can't recall not having fun looking thru scopes.) :)
Perhaps some personal background won't be too out of place here. I recall
going to work years ago at an optics shop and having an 8" Newtonian I had
labored with to make 1/10 wave I must say I was a bit dismayed when my boss
proclaimed it a smooth 1/2 wavefront "not too bad". Indeed today I think he
was being perhaps generous but it still provided some fine views and I was
not ashamed to set it up next to comparbly sized commercial scopes of that
era. One result is that while I may hazard a guess as to wavefront quality,
I'll be pretty vague most of the time as it really does take considerable
effort and equipment to quantify the whole optical surface across several
diameters at the same time. Struggling through a time a decade or two ago
when 'diffraction limited' was the big ticket, I can certainly understand
folks who are reluctant to accept anything without a numbered certification
attached. That term was as abused as it is meaningless when not defined. I
null test my optics but this is another term that probably is meant more to
impress than explain though there are aspects of it that are comfortable for
me as an aspiring optician when I'm using it within it's limitations and
strengths.
To try to come to some point in this long post, it would be that optics
production is tedious, but more and more opticians are becoming very
capable indeed of making marvelous scopes for the amateur market. There are
clear advantages to interferometry but time and setup can escalate the cost
of the finished product. An optician's skill will still determine the final
product though an accurate testing procedure may help him develop that skill
in a less chaotic manner. I'm in the fortunate position of making optics as
a hobby so time is not such a consideration but that has resulted in my
being willing to stumble around with less efficiency and I don't produce
optics at a very high rate (in fact it is downright embarressing how long
some of them take <G>). Others have refined their methods that leave me with
a tinge of envy but they have valid commercial interests to attend to (like
putting food on their table). I only hope that the users of these optics
will appreciate the values they receive today and not be confined by the
limitations -perceived or real- but will explore with wonder and vitality to
the full potential of whatever equipment they acquire. I must say I've
enjoyed and learned much from these recent threads. It just seems we
sometimes document scopes and sometimes scope builders and forget to view.
Tonight if you set up your best scope right next to mine in my backyard,
they will provide the exact same quality image and I'll stand by that. (It's
raining, BTW) Tomorrow night may have differing conditions so it would be
invalid to compare then of course <grin>.
Thanks,
Mike Spooner
Sue_and_Alan_French <Sue_and_A...@email.msn.com> wrote in message
news:u9AVUmHf$GA.191@cpmsnbbsa03...
> Hi Mike,
>
> People should read "Optical Quality in Telescopes" in the March, 1992
issue
> of Sky & Telescope, and pay close attention to what Douglas George and
> Terence Dickinson say about the differences between a true 1/4 wave and
true
> 1/10 wave telescope.
>
> Clear skies, Alan
>
Zane
>I'm new to this discussion but I've read all the postings. I too have
>wondered about these incredible numbers. I run an (US Navy) optical
>fabrication facility. My primary optical metrology is phase
>interferometry.
(snip)
Hello Randy
Can an MTF for something like a f/5 or f/6 12" really good mirror be
directly and accurately measured with the equipment you guys have? Do you
know of other people who can do this?
Zane
randy_...@my-deja.com
Hi Zane
I can't do this directly - I do not have a MTF test setup in my lab. My
Wyko software will calculate the MTF based on the phase infomation. Or I
can do this in Zemax using the Zernike term fit of the wavefront
provided by the software. Within the intrinsic accuracy of the
interferometric measurement this is valid. As far as actually doing a
MTF measurement I don't see why it couldn't be done with the commercial
equipment available. I would think the difficulties encountered would be
as great or greater than interferometry since you have to come up with
auxillary optics, deal with environmental issues and so on. We do
conduct MTF tests of IR systems but that is done on a range with large
fixed sources.
Randy
todd_...@my-deja.com
off thread for a year or so you will note that I settled on taking the
numbers, ie 1/20th wave, and have "assumed" that on the interferometer
would test them at 1/5th to 1/10th wave (1/2 to 1/4 of the 1/20) based
on discussion on this group, my own observations, and observations from
Richard Whalen. This tentatively confirms my "assumption" and lends some
usefulness to the numbers.
I also would like to note that Galaxy did describe this effect to me
when I was shopping for mirrors. They hit the issue right on.
Todd Gross
> Hi Roland,
>
> a great idea to do such comparation, so my Portaball have an real
> chance to use a real ~ 1/8 wave mirror, which is more than I expected
> during placing my order.
> great Post, thanks
>
> Markus
>
> chri...@aol.com (Chris1011) wrote:
> > >> I would like to know at what
> > level of performance these test out, so that I might be enlightened
> as to what
> > specific wavefront error can be had and still provide the mysterious
> and
> > elusive 'diffraction limited' performance that>>
> >
> > I was curious to see how interferometry compares to the focault
test,
> so i did
> Furthermore,
> > the interferometer measures astigmatism and residual coma, which the
> Focault
> > test does not do, at least not without an ungodly amount of work.
> >
> > As an example, I took one mirror which has the following
> interferometrical
> > data: Astigmatism=.056, Coma=.039, Spherical=.042, all in waves P-V.
> Overall
> > P-V=.080, or about 1/12 P-V.
> > With the Focault test, only the spherical component would be
> measured, so the
> > mirror would come out 1/23 wave P-V. Actually, at this focal ratio,
> I could
> > not read the knife edge accurately enough to be sure of this level
of
> accuracy.
> > In fact, it looked better than 1/23 with the knife edge.
> >
> > My conclusion is that the interferometer values are somewhat overly
> pessimistic
> > and the knife edge is somewhat overly optimistic on mirror quality,
> and that
> > half way in between for each is probably closer to the truth. I
would
> say that
> > a 1/30 wave Focault tested mirror will probably test on the
> interferometer
> > between 1/10 and 1/14 P-V., a not too shabby mirror. A close reading
> of
> > Malakara shows that this is the case. He does not favor the use of
P-
> V at all,
> > and says that RMS readings are more representative of actual
> performance. A
> > mirror with 1/14 wave RMS would be just barely acceptable, and 1/50
> RMS would
> > be pretty much perfect.
> >
> > Roland Christen
> > ASTRO-PHYSICS
> >
>
to...@shore.net
> over the map on the same subject. I think we do get right back to the
best
> startest acceptable to the scope owner for the most reasonable amount
of
> money. This current "spec" war thus seems a little silly. As for the
> upcoming "face-offs", Todd Gross has already done this. The 18" dob
won :-).
>
thanks. Actually, additional face-offs are always helpful, especially
since I rarely have good seeing. I was blessed with good seeing this
morning allowing me to test the 10" Zambuto Teleport right up against
the revered FC125. Both star tests are outstanding. On Jupiter, Saturn,
and with excellent seeing on the Moon, aperture won as would be
expected. I have to admit the contrast between the planet and the sky
was very impressive in the 5"
randy_...@my-deja.com
P.S. There is no real reason to test the MTF of a system that is close
to diffraction limit as there are other test methods that can be use.
MTF testing is valuable for systems that have a lot of aberration such
as camera lenses. It would be, nonetheless, interesting to do such a
test as a comparison. Off hand, I can't direct you to someone that can
do this except the large firms and some testing services. These are
hideously expensive.
Randy
Ric Rokosz
Organization: Toronto Free-Net
Distribution:
Mike Spooner (spo...@page.az.net) wrote:
: null test my optics but this is another term that probably is meant more to
: impress than explain though there are aspects of it that are comfortable for
: me as an aspiring optician when I'm using it within it's limitations and
: strengths.
Mike
Could you expand on the issue of "within it's limitations and strengths"
of the nulling aspect in testing?
Ric
Mike Spooner
Mostly that it is difficult to assign meaningful numbers to the results. Of
course the zonal differences can be measured but by the time I get as close
to a null as I can figure, I wouldn't put much stock in my ability to get
precision readings. At times I'm not sure if I'm testing the mirror or the
reference flat with respect to subtle zones and so the reference flat itself
is a source of limitation. Of course precision flats can be manufactured in
the old 1,2,3 method using interference to a very high accuracy.
Astigmatism is another problem to identify that can be frustrating to see in
it's more subtle forms. Overall though, I really like using autocollimation
as I'm not adept with Foucault testing and the ease of seeing gross errors
in the doublepass is comforting for me. With a Foucault, there is the matter
of data reduction and seeing how it compares to whatever standard curve is
being sought - with the reduction programs discussed on the ATM group (as
you're probably aware) this is much less work than in the past. But the null
test gives a visual of the error, it's approximate area of involvement, and
with a bit of experience, an indication of where the work needs
concentrate - all at a fairly quick glance. Of course some of this is bred
by familiarity so I'm certain to be biased to a degree. I still like to star
test mirrors in solid seeing as a check of the final figure, as I found I
can let optimism, fatigue or impatience prejudice my assessment of the bench
tests. Since I haven't yet made a perfect optic <g> it is always a struggle
to detemine what errors there are, how or if they'll affect the image, and
if I can make it better in a reasonable amount of time or if I need to leave
it alone. Now if the null test could answer those, I'd say it wasn't as
limited but I fear I'll always be the biggest limitation no matter what test
is used. :)
--Mike
Ric Rokosz <bz...@torfree.net> wrote in message
news:FqC34z.I1...@torfree.net...
Ron Wodaski
contrast was. I watched two stars emerge from behind the 95% full moon's
bright edge: first a mag 5.2 star (popped right into view), and then a mag
7.5 star (took me a moment, but once I found it, the star was clearly
visible). There was some high-altitude haziness and some thin cirrus here
and there, so I have no idea just how dim a star I could observe yet. Still,
these results are very encouraging. I could hardly see any stars near the
moon with my SCT or other non-refractors. I've been seriously impressed with
the contrast in the FC-125.
Ron Wodaski
<to...@shore.net> wrote in message news:88u0fj$r51$1...@nnrp1.deja.com...
Zane
(snip)
>P.S. There is no real reason to test the MTF of a system that is close
>to diffraction limit as there are other test methods that can be use.
(snip)
As this discussion points out, the interpretation and artifacts of
interferometric data at the surface qualities being mentioned can be less
than definitive in arriving at an MTF. Since the MTF defines all the
things being argued about, primarily planetary performance, a direct
measurement would be good.
I know that measuring the PSF or LSF defines the MTF, but there are still
signal-to-noise and sampling issues. The best "machine" I've personally
seen only had 8" diameter, 1/10 wave collimation optics and had to be
analytically corrected for variation of sampling with sampling apertures
getting to be on the order of the light wavelength. (Even at this level,
measurements had to be made with no humans and very little electronics in a
special room.)
Such measurements, if possible, would sure settle a host of arguments.
Zane
Bryan Greer
> Where in the world do you find books like this
> Bryan? You don't buy them, like I do, do you?
>
I'm sure my library pales by comparison.
> (excerpt from book)
> "The same surface can have many different
> values for the rms wavefront roughness depending
> on the instruments used to make the measurements."
>
This is a little scarey...I almost quoted the same passage. :-)
> These are surface roughness measurements
> (resolution down on the order of 1A RMS),
> and while I would love to have a Newtonian
> mirror with such smoothness, we are
> well away from the day that amateur mirrors
> are this smooth.
>
I agree, but it is a little surprising to me just how smooth the surfaces can be
with even a modest amount of attention to detail. Here are some numbers on some
recent samples I sent out for atomic force microscope (AFM) scanning (testing by
Charles Evans & Associates; Sunnyvale, CA). These are samples of my Pyrex and
fused silica secondary mirrors, commercially polished with red rouge. The
optician knew I would be AFM scanning these, so you could argue he took his time.
However, I've randomly sampled other pieces in the past with somewhat similar
results.
sample size: 512 x 512 nm
sampled points: 512 x 512 (262,144 total)
scale: 1 nm/pixel
vertical resolution: 0.5 angstroms
Sample rms (angstroms)
----------------------------
FS #1 2.7
FS #2 2.9
Pyrex #1 4.9
Pyrex #2 5.4
I'm working on a 8.75" quartz primary mirror right now for my personal use. Just
for kicks, I'm going to use ball milled rouge to see how smooth it turns out. I
can't wait to see the AFM data from it.
> Point well taken. However, these threads about
> optical wavefront quality are starting to give the
> impression that the interferometer is a flawed tool.
>
Yes, that's why I asked Roland about his conclusion that the interferometer
indicating an "overly pessimistic" value.
> S.A.? Wavefront yes, but S.A.?
>
Yep, SA. It's not uncommon for high-end APOs to perform at this level under real
conditions in the field.
I gotta go...they're refinishing our wood floors today, and I'm about to be
"painted in"!
Sandy McNamara
<spo...@page.az.net> writes:
>I only hope that the users of these optics
>will appreciate the values they receive today and not be confined by the
>limitations -perceived or real- but will explore with wonder and vitality to
>the full potential of whatever equipment they acquire.
What a wonderful thought, Mike! I do hope others are listening.
> I must say I've enjoyed and learned much from these recent threads.
Yes, indeed. I'm never going to start grinding my own mirrors but I appreciate
being a knowledgeable user. The rational discussions and explanations of the
merits, faults, and inherent differences between different methods of assesing
optical quality has been extremely interesting. It has been nice to get away
from the "my way of measuring is best - my numbers are different from yours -
therefore you must be wrong" bashings that were going on.
Now if winter would only hurry on its way, we can all get out and enjoy them in
the peaceful pursuits for which they were designed :-)
Cheers,
Sandy Mc.
Sandy...@aol.com (Sandy McNamara)
40d 26' N 89d 13' W
TMBack
>I'm sure my library pales by comparison.
It was just a joke Bryan, I was impressed that
you had this rather rare book.
>This is a little scarey...I almost quoted the same
>passage. :-)
Well, it was a good passage. :-)
>I agree, but it is a little surprising to me just how
>smooth the surfaces can be with even a modest amount
>of attention to detail. Here are some numbers on some
>recent samples I sent out for atomic force microscope
>(AFM) scanning (testing by Charles Evans &
>Associates; Sunnyvale, CA). These are samples of my
>Pyrex and fused silica secondary mirrors, commercially
>polished with red rouge.
Red rouge is still the ticket!
>The optician knew I would be AFM scanning these, so
>you could argue he took his time. However, I've
>randomly sampled other pieces in the past with
>somewhat similar results.
>
>sample size: 512 x 512 nm
>sampled points: 512 x 512 (262,144 total)
>scale: 1 nm/pixel
>vertical resolution: 0.5 angstroms
>
>Sample rms (angstroms)
>----------------------------
>FS #1 2.7
>FS #2 2.9
>Pyrex #1 4.9
>Pyrex #2 5.4
Woowha! I love them fused silica flats. You got
to make me one (you know, the big one I was asking
about) for my next BIG telescope project.
>I'm working on a 8.75" quartz primary mirror right now
>for my personal use. Just for kicks, I'm going to use ball
>milled rouge to see how smooth it turns out.
>I can't wait to see the AFM data from it.
I can't either. You will bring this beauty to Astrofest,
won't you?
>Yes, that's why I asked Roland about his conclusion
>that the interferometer indicating an "overly pessimistic"
>value.
That puzzled me also.
>>S.A.? Wavefront yes, but S.A.?
>
>Yep, SA. It's not uncommon for high-end APOs to
>perform at this level under real conditions in the field.
I think you're still confusing S.A. with wavefront.
You wrote:
>> If I had my way, peak-to-valley numbers would be
banished from the planet. Our eye/brain perceives RMS
(or equivalently Strehl, MTF, EER, etc.) -- not P-V.
Sincerely,
Bryan Greer <<
You were right the first time. The eye/brain perceives
not the P-V (or the SA3 in this case), but the entire
wavefront. I have stacks of interferograms that have S.A.
numbers with vanishingly small values, but the overall RMS
wavefront is realistic, and with a very high Strehl.
Thomas Back
TMB Optical
Pete Johnson
> >>The term 'crappy optics' was my summary of the rather obvious
> thrust of some of the posters (but not you) had taken as to the 'amateur'
> mirrors
> from Carl and John and others. >>
>
> Not one person has said the optics are crappy, NOT ONE!
>
> What we on the front lines resent is a quality claim of 1/28 P-V wavefront
> error which cannot be supported by the Focault test. All the Wow factor in the
> world won't change that. I'm sure that the same Wow will be heard from anyone
> getting a mirror from any other quality mirror maker whose measurements with
> interferometry may show ONLY 1/8 to 1/10 wave P-V. Has anyone done a comparison
> between a 1/8 wave interferometer measured Ceravolo mirror and a 1/28 wave
> Zambuto mirror?
>
> Roland Christen
Roland,
I have a Swayze mirror, 24" which I am very pleased with. Swayze does not
quote misleading numbers on his mirrors but rather just guarantees that
you will be pleased or he will fix the problem. I think that is a very
honest way of doing business, not that others are dishonest :-). Swayze
says he used a Ross Null test using a certified lense from Ceravolo (your
post made me think of this). How does a Ross Null test compare to
interferometer or Focault testing? One does not here of many mirror makers
using it.
Thanks in advance
Pete Johnson
Ron Wodaski
here, but I haven't seen much about contrast transfer. I recently acquired a
top-quality fluorite refractor, and I noticed right away how stunning the
contrast is compared to any other scope I've ever looked through (including
what I'll call medium-quality fluorite refractors). I've been reading
"Telescope Optics" by Ruten and van Venrooij, and they make a strong case
for measuring contrast trasnfer _and_ optical quality in order to really
understand the performance of a given scope.
How is contrast transfer measured? Why don't we see numbers on contrast
transfer, since it would seem to be able to help out when assessing scopes
for various intended uses? In particular, high contrast transfer (that is,
transfer from the source to the image) is a great help in planetary
observing.
Ron Wodaski
"Chris1011" <chri...@aol.com> wrote in message
news:20000219134219...@ng-ca1.aol.com...
> >>
> I understand your point about how the Foucault test misses four of the
five
> Seidel
> aberrations. The undersampling of the Foucault test, along with operator
> variability, are the two main sources of error. However, I don't quite
see why
> you
> came to the conclusion that an interferometric test is overly
"pessimistic".
> Are
> you saying this is a general truism, or just in this particular case? >>
>
> In the case of Perkin-Elmer, I would assume that they have a very
expensive
> system costing in excess of $1 mil, which does phase measurement and can
be
> accurate to perhaps 1/100 wave. This is not the case with my "cheap" $50K
> system. I have characterized my interferometer's errors to be on the order
of
> 1/25 wave. However, it is very useful in showing where the corrections
need to
> be made, and shows astigmatism very easily. I use it as a quality control
> device, not as an absolute measurement. When the optic under test is
better
> than 1/10 wave including the interferometer errors, I can stop messing wit
h it
> and get it ready for sale.
>
> In the case of interferometer P-V measurements, every little departure
from
> perfection adds to the error, even if it has little or no effect on the
final
> image. The RMS measurement takes this into account, and reduces the effect
of
> small localized errors on the final number. Unfortunately, this is not
easily
> computed with the Foucault test.
>
> Roland Christen
> ASTRO-PHYSICS
Chris1011
interferometer or Focault testing? One does not here of many mirror makers
using it.>>
Could be quite accurate. I know someone in England who could achieve 1/8 wave
P-V with that method. Using a high quality flat in autocollimation is more
accurate since all errors are multiplied by two.
Roland Christen
Chris1011
top-quality fluorite refractor, and I noticed right away how stunning the
contrast is compared to any other scope I've ever looked through (including
what I'll call medium-quality fluorite refractors).>>
I think the type of contrast you are referring to in your refractor has less to
do with the accuracy of the figure than the lack of low intensity scatter that
pretty much all mirror systems suffer from. Even when there is dust on an
objective lens, there is little or no forward scatter (the part that ends up in
the eyepiece), yet there is considerable back scatter, back out to the sky. In
a mirror system, on top of the scatter inherent in aluminum coatings, the
smallest bit of dust or airborn debris on the mirror surface radically
increases the backscatter (in a mirror thats what ends up in your eyepiece).
Therefore, areas adjacent to bright objects, like the edge of the Moon would be
brighter and thus hide things like faint stars.
This contrast effect is one reason why coronagraphs have refractive optics,
rather than mirrors.
Roland Christen
ASTRO-PHYSICS
Bryan Greer
> It was just a joke Bryan, I was impressed
> that you had this rather rare book.
>
Actually, it isn't so rare anymore. The Optical Society of America
reprinted a few years ago as an inexpensive paperback, and it still has
the color plates in it. If I remember right, it was about $35 when I
bought it. It is a good reference source, and I go through spells where
I refer to it often. No optical geek should be without it!
(For those coming in late on this thread, we're discussing the book
"Introduction to Surface Roughness and Scattering", by Jean Bennett and
Lars Mattsson; ISBN 1-55752-108-5.)
> I can't either. You will bring this beauty to Astrofest,
> won't you?
>
It depends on how well it turns out.<g> Seriously, I'm in the process
of setting up a new little optical shop, so there is no glass action
until that's finished. The new mirror may not be finished by this
year's Astrofest (though I will bring my whiz bang thermal demonstration
rig). Heck, if my new scope really performs, I'll drive it the 1-1/2
hour up to you! A little advance good press from TMBack should propel
it to stardom even before its debut!
> I think you're still confusing S.A. with wave front.
>
Yes, yes, yes, yes........ NOW I see what you are saying (Don't be so
obtuse, Thomas. Speak up, man!<g>). Yes, I am guilty of using the
dreaded SA P-V measurement--the same thing I just said I wanted banished
from the planet. I was just using the SA value since that is what most
people (rightly or wrongly) use when they evaluate a scope, and it
assumes no other serious flaws (the same often-wrong assumption that
lets you convert P-V to rms by simply multiplying by some factor). It
was an off-the-cuff remark, and the main point made was that super
accurate optics are not necessary. As we both agree, the eye perceives
the MTF of an optical system--not P-V.
Ron Wodaski
others (while I am easier to baffle than most. <g>).
But that begs the question: why don't we see contrast transfer quoted more
often? If a designer has done a good job with baffling, optical design, and
keeping obstruction size down, one can measure contrast, but I just never
see numbers of this. Have you seen the chapter I'm talking about in
"Telescope Optics?" It's quite interesting, pages 215-223. The formula for
contrast transfer:
I1 - I2
C = -----------
I1 + I2
where I1 and I2 are the intensities of adjacent surfaces, and I1 > I2.
"...contrast transfer is measured by placing a grating having a sinusodial
intensity distribution as in object in front of the optical system, then
measuring the contrast of the resultant image. The ratio between image
contrast and object contrast is called the contrast transfer coefficient."
(p. 217)
However good contrast is achieved (choice of telescope type, baffling,
surface smoothness, etc.) it would be ever so useful to be able to quantify
it routinely.
Ron Wodaski
"Chris1011" <chri...@aol.com> wrote in message
news:20000222183044...@ng-cf1.aol.com...
Mike Spooner
Here's something I often wondered about when the doublepass is mentioned. I
understand about the light traversing the error twice and it's effect at the
tester, etc., but no one mentions that the optic under test is now working
at half the f/ratio of the same optic tested at C of C. How does sensitivity
of knife-edge tests with respect to f/ratio affect this whole relationship?
As you may have seen, I use and like autocollimation with a good flat but I
still haven't determined how much the doublepass really means. Just one of
those off the wall thoguhts I've had. <g>
Thanks,
--Mike Spooner
Chris1011 <chri...@aol.com> wrote in message
news:20000222182145...@ng-cf1.aol.com...
Roy Diffrient
>
Snip!!
>I had a chance to look through the scopes of the S & T test at Stellafane a
>couple of years ago but the daylight seeing over the ground was not the best
>and the thermal heat problems were at least bad enough to overwhelm the
>testing (clouds prematurely ended my attempts). I couldn't stay for the
>night due to a tight timetable. :(
Hi Frank,
I was at Stellafane then and took that test, too! There was a very good reason
for all those thermal heat problems, and it wasn't just the daylight seeing:
They had set up the scopes so that they were looking at an artificial star
(and later Polaris) right over the gigantic charcoal-fired barbeque at the food
tent!! It was not much better at night, so you didn't miss much, IMO :-<
Roy Diffrient
Monkton, MD
bratis...@my-deja.com
"Mike Spooner" <spo...@page.az.net> wrote:
> How does sensitivity
> of knife-edge tests with respect to f/ratio affect this whole
relationship?
> As you may have seen, I use and like autocollimation with a good flat
but I
> still haven't determined how much the doublepass really means. Just
one of
> those off the wall thoguhts I've had. <g>
Certainly some valid wall thoughts you have, Mike :-)
The sensitivity of the knife edge test varies inveresly with the square
of the focal or radius ratio. So, null testing with double pass at half
f/ doesn't double the errors (in fact it DOES, but they are four times
harder to see :-)
Ross null will also reduce f/ratio at knife, as will Waineo sphere.
Of all null tests, Waineo is most favourable as far as sensitivity at
knife edge goes. To obtain same sensitivity as double pass
autocollimator one needs to test at 1.414 of scope's final f/ratio (or
0.707R). This calls for test sphere of about half size of to be tested
paraboloid; certainly much cheaper and easier to make than a large flat.
With larger sphere you may test at even longer final f#.
Of course, all bets are off when you make all spherical optics, or some
weirdo aspheres that can't be nulled using Waineo sphere (and they won't
be nulled by Ross lens either). That's why large flats are so desirable
(and expensive) !
Bratislav
bratis...@my-deja.com
chri...@aol.com (Chris1011) wrote:
> the lack of low intensity scatter that
> pretty much all mirror systems suffer from. Even when there is dust on
an
> objective lens, there is little or no forward scatter (the part that
ends up in
> the eyepiece), yet there is considerable back scatter, back out to the
sky. In
> a mirror system, on top of the scatter inherent in aluminum coatings,
the
> smallest bit of dust or airborn debris on the mirror surface radically
> increases the backscatter (in a mirror thats what ends up in your
eyepiece).
Not just that, if we agree that optical glass can be polished to the
same RMS level a mirror substrate can (quite close assumption), the same
amount of microripple will cause at least TEN TIMES more scatter on a
mirror than it will on a lens.
For a given surface error, wavefront will be deflected almost four times
by reflection (using BK7 here); squared to take area into account that
is ~15 times ; multiplied by sqrt(2) to account for two surfaces of the
lens we arrive at almost exactly ten ! :-)
And this didn't even take into account dreaded aluminium layer. For that
eye hurting limb contrast nothing comes close to a good refractor.
But lets not forget that this low level scatter is mostly aesthetic, and
its effect is mainly to drop MTF curve overall a notch. Nothing which a
healthy dose of a larger aperture can't fix ;-)
Dan Chaffee
<spo...@page.az.net> wrote:
>Perhaps some personal background won't be too out of place here. I recall
>going to work years ago at an optics shop and having an 8" Newtonian I had
>labored with to make 1/10 wave I must say I was a bit dismayed when my boss
>proclaimed it a smooth 1/2 wavefront "not too bad". Indeed today I think he
>was being perhaps generous but it still provided some fine views and I was
>not ashamed to set it up next to comparbly sized commercial scopes of that
>era.
Nice post, Mike.
Tell me though, what was the max. mag./inch-aperture that this 8"
scope held up to? And BTW, was he more specific about the kind
of 1/2WF aberration?
Dan Chaffee
JMc
Thanks for these data on Pyrex vs. quartz. I think we now see the real
reason AP is using quartz/FS mirrors in their new MCTs.
Jim McSheehy
> Bryan Greer wrote:
> I agree, but it is a little surprising to me just how smooth the surfaces can be
> with even a modest amount of attention to detail. Here are some numbers on some
> recent samples I sent out for atomic force microscope (AFM) scanning (testing by
> Charles Evans & Associates; Sunnyvale, CA). These are samples of my Pyrex and
> fused silica secondary mirrors, commercially polished with red rouge. The
Ric Rokosz
: >> How does a Ross Null test compare to
: interferometer or Focault testing? One does not here of many mirror makers
: using it.>>
: Could be quite accurate. I know someone in England who could achieve 1/8 wave
: P-V with that method. Using a high quality flat in autocollimation is more
: accurate since all errors are multiplied by two.
Using the Ross depends on the focal lenght of the lens - the longer the
better and it seems to equal the autocollmination.I use a 12 inch
diameter lens with 91 inch focal lenght.Also use a 12 inch flat and the
images on the null are very close.
Ric
the...@frontiernet.net
I had William Bell make me a Ross null test slide, but their
instructions are incomplete.
Thanks
Bob
Mike Barrs
uncoated mirrors?
I once read somewhere (maybe it was on the ATM list) that although quartz
can take a finer polish than any other material, the difference (vs. Pyrex,
Zerodur, or other materials) would be lost after a mirror was aluminized. In
other words, the surface roughness of the coating would be significantly
higher than the best surface attainable on quartz. So the reasons to use
quartz would seem to have more to do with thermal stability, and not the
surface finish.
Does anybody here know if this is true, or is it just an ATM myth? Maybe it
depends on the quality of coating used? Is there a special reflective
coating that preserves the surface smoothness of quartz?
Mike Barrs
JMc <tri...@my-deja.com> wrote in message
news:38B3DC65...@my-deja.com...
> Bryan,
>
> Thanks for these data on Pyrex vs. quartz. I think we now see
> the real reason AP is using quartz/FS mirrors in their new MCTs.
>
> Jim McSheehy
>
> > Bryan Greer wrote:
> > I agree, but it is a little surprising to me just how smooth the
> > surfaces can be with even a modest amount of attention to detail. Here
> > are some numbers on some ecent samples I sent out for atomic force
> > microscope (AFM) scanning (testing by Charles Evans & Associates;
> > Sunnyvale, CA). These are samples of my Pyrex and fused silica
secondary
> > mirrors, commercially polished with red rouge. The ptician knew I would
Mike Spooner
recall trying it at higher powers (didn't have a barlow at the time). It had
smooth undercorrection that I remember clearly showing in autocollimation
before I refigured a few years ago. The most memorable view was low power of
the Horsehead Neb 20 years ago in some amazing skies near Flagstaff, Az. No
filter but it looked like a tiny photo! Saw it in a C-8 that friend had
close by also. Haven't seen it since. :( Skies and eyes really do make a
difference.
At 200x I saw several small spots in the SEB but only one night of excellent
seeing. Vaguely recall the GRS as somewhat difficult. Saturn and Cassini
div. looked pretty good IIRC but not like I'm used to now. Mars I distinctly
remember as always being a struggle to more than see the polar caps and some
light shading and I always considered it a low contrast object. Now I
consider it as having some of the higher contrast detail of any of the
planets! Quite a switch, huh?
At anyrate I refigured it a few years ago and it'll take 400x in a ho-hum
manner (and Mars looks a lot better).
--Mike
Dan Chaffee <dcha...@gvi.net> wrote in message
news:38b37d1b...@news.gvi.net...
Bryan Greer
> Are those AFM scanning test numbers from coated mirrors, or polished but
> uncoated mirrors?
Good question; these numbers are for uncoated surfaces.
> I once read somewhere (maybe it was on the ATM list) that although quartz
> can take a finer polish than any other material, the difference (vs. Pyrex,
> Zerodur, or other materials) would be lost after a mirror was aluminized.
Well, maybe and maybe not. I honestly don't know how much of the smoothness
advantage is preserved after coatings are applied (but I'm in the process now of
finding out). In theory, if you start with a smoother surface before coatings,
you'll end up better off after the coating is applied. This is because the
migrating ions preferentially attract to the high peaks. Thus, thin films
exaggerate any roughness already on the substrate--at least for a vacuum
deposition process. If the coating is thick enough, this snowballs into some
pretty ugly columnar structures that scatter a lot of light. For a given
material, the magnitude of this exaggeration is almost purely a function of how
thick you apply the coating. Thicker is obviously rougher.
I also don't know just how large the microroughness can be before it becomes
noticeable to the eye as an increase in veiling glare. It seems everybody has
their own personal theory on this (including me), but I've never seen a test to
conclusively determine how much microroughness is acceptable. This is really
the $1M question. Obviously, there's no point in beating your wallet up to get
a 5 angstrom rms final surface finish if your eye can't see the difference from
a 25 angstrom rms finish.
> So the reasons to use
> quartz would seem to have more to do with thermal stability, and not the
> surface finish.
Yes, there are a few good reasons for using quartz (and other materials) as a
substrate. Strictly from a manufacturing point of view, quartz is nice since
you can read the pieces more accurately on the block. This gives the optician a
better idea of what he/she has before calling the block finished. Presumably,
this results in slightly better pieces.
> Does anybody here know if this is true, or is it just an ATM myth? Maybe it
> depends on the quality of coating used? Is there a special reflective
> coating that preserves the surface smoothness of quartz?
Most dielectric films have less microroughness than metallic films (though SiO2
can end up quite rough if not applied with the right equipment and process).
However, metallic films can be "peened down" to extremely smooth surfaces if you
propel energetic ions at the coating during or after it's applied (i.e., ion
assisted deposition (IAD)).
bratis...@my-deja.com
> multiplied by sqrt(2) to account for two surfaces of the
> lens
should be "divided by sqrt(2)", sorry
bratis...@my-deja.com
bz...@torfree.net (Ric Rokosz) wrote:
> Using the Ross depends on the focal lenght of the lens - the longer
the
> better and it seems to equal the autocollmination.I use a 12 inch
> diameter lens with 91 inch focal lenght. Also use a 12 inch flat and
the
> images on the null are very close.
Autocollimation with a flat doubles the errors, but K-E contrast drops
with square of the final f ratio. This means if your Ross lens is long
enough not to shorten final f/ratio too much, not below 0.7R/D
(autocollimation test is at F/D = 0.5R/D), the sensitivity at knife edge
will be the same.
12" Ross lens !? Wow !
Bratislav
David Smith
> see numbers of this. Have you seen the chapter I'm talking about in
> "Telescope Optics?" It's quite interesting, pages 215-223. The formula for
> contrast transfer: ...
>
> "...contrast transfer is measured by placing a grating having a sinusodial
> intensity distribution as in object in front of the optical system, then
> measuring the contrast of the resultant image. The ratio between image
> contrast and object contrast is called the contrast transfer coefficient."
> (p. 217)
Hmm. In literature I have read, the Contrast Transfer Function was defined in
terms of square-wave intensity distribution, as contrasted (I would like to
claim No Pun Intended, but you wouldn't believe me) to the Modulation Transfer
Function, which is defined for sinusoidal distribution. The authors pointed
out that bar (Ronchi) gratings are easier to produce than sinusoidal, and that
for fine detail, there is not much difference, as they only differ in 3rd and
higher harmonics.
But, CTF and MTF are functions of spatial frequency and don't give you a single
number to bandy about.
David Smith
Ron Wodaski
you calculate the ratio of source contrast and image contrast.
Ron
"David Smith" <d...@hpl.hp.com> wrote in message
news:38B46562...@hpl.hp.com...
Nils Olof Carlin
Roland Christen said:
"In the case of interferometer P-V measurements, every little departure
from
perfection adds to the error, even if it has little or no effect on the
final
image. The RMS measurement takes this into account, and reduces the effect
of
small localized errors on the final number. Unfortunately, this is not
easily
computed with the Foucault test."
There is one (that I know of) evaluation program that does this - James
Burrows' SIXTESTS
( it runs under DOS and is available for download at
ftp://ftp.halcyon.com/pub/users/burrjaw/sixtests.zip ).
Unlike the more primitive algorithm of Texereau (and computerized versions -
see "How to make a Telescope"), it interpolates the data points and
reconstructs the best fitting parabola, and computes the RMS and Strehl (as
well as P-V for "backward" compatibility ;-).
Of course it also plots the estimated error curve.
"Focusing" by finding the best reference parabola is of course necessary to
get Strehl/RMS, but not practical in a "hand" algorithm such as Texereau's.
BTW the name SIXTESTS refers to the fact that it evaluates Foucault, Gaviola
and "Poor Man's Caustic" readings, for fixed or moving source, 6
combinations. Incidentally, this may for all I know be the only available
way to evaluate the PMC correctly, as the old "divide the KE offsets by 3
and evaluate as Foucault" gives grossly incorrect estimates of the wavefront
error. SIXTESTS numerically solves the appropriate differential equation,
see its documentation.
What if you could re-run your Foucault data through SIXTESTS and compare the
RMS?
The limited sampling of Foucault (as compared to interferometry) is a
limitation that an evaluation algorithm cannot circumvent, of course, but I
am convinced SIXTESTS makes the best possible use of the Foucault readings
obtained AND it sure helps to fulfill Bryan Greer's wish that I share (only
I would extend it beyond "the planet" to any planet visible from Earth ;-)
"If I had my way, peak-to-valley numbers would be banished from the planet.
Our eye/brain perceives RMS (or equivalently Strehl, MTF, EER, etc.) -- not
P-V."
I would say that what we perceive most directly looking at a star is the
point spread function PSF ;-) I think it would deserve wider use - from it,
the others can be derived.
(actually, SIXTESTS calculates the PSF)
In a recent thread on ATML, a simulated mirror with a 5th order spherical
aberration of 1/8 wave was used to evaluate SIXTESTS. SIXTESTS reported a
P-V error of 1/4.6 waves!(at best focus, the error curve was the spitting
image of a 6th order Zernike polynomial). This rather striking difference is
due to the fact that the P-V error is NOT in general minimum at best focus
(where the RMS error is by definition minimum, if higher terms can be
ignored).
I also would like to offer another explanation of why the RMS and Strehl
correlate better to what we see. In a perfect telescope, at focus, all parts
of the aperture contribute in phase to build the Airy "diffraction peak" -
and the height of this peak is the "denominator" of the Strehl ratio.
In a real telescope, any part of the aperture causes an optical pathlength
error and a phase error phi (wrt the best reference wavefront). The
contribution at focus will now be less than 1, and instead proportional to
cos(phi). cos(phi) can be written 1-0.5(phi)^2... where higher terms can be
ignored if the phase error is small. The total lowering of the peak (and
thus the Strehl ratio) is proportional to the mean square deviation over the
aperture(not strictly the square root of it=RMS, but that is trivial).
"Mike" said:
"....Why do interferometrical test reports often show what I think are
inconsistent values for ptv, RMS and strehl ratios? Look at the TMB
105/650 on APM`s site. It shows:
ptv 0,226
RMS 0,027
Strehl 0,972
How can a lens that bearly makes the diffraction limit have a Strehl of
97."
This will be confusing to anyone who thinks the P-V error can define
"diffraction limited". A mirror with pure low-order spherical aberration of
1/4 wavelength P-V will have a RMS error of 1/13.4 (IIRC) and a Strehl ratio
of 0.8, and I believe it was of such an example that lord Rayleigh said "the
performance begins to be decidedly prejudicial".
This criterion of "diffraction limited" can be used with Strehl >0.8 or
RMS<1/13.7 wave, but NOT with P-V<1/4 wave in the general case. SIXTESTS
will often report good enough RMS/Strehl when the P-V is worse than 1/4
wave. Let us support Bryan in banishing P-V ;-)
SIXTESTS will also often confuse anyone who thinks "diffraction limited" is
related to P-V measurements.